This post is a specific response to Tom Nolle’s recent blog: What Went Wrong with NFV: The Operator View.
I provide a perspective on the points Tom raised, based on my experience leading business model strategy / transformation at a leading Networking supplier, Juniper Networks. Tom’s blog reflects the operator view as articulated by Tom, and not necessarily Tom’s own views, according Tom’s comments in the post. The purpose of this post is not to critique Tom’s post, but to provide experience-based perspective.
- Most transformation efforts fail in most companies, regardless of industry.
- There are many common reasons for transformation failure that are independent of whatever challenges NFV has experienced. Some of what NFV has experienced is simply a manifestation of what is common in transformation efforts.
- The common experience with transformation failure should lead to an increased focus on inorganic strategies.
- Nonetheless, there are specific NFV observations worth making: supplier CFOs manage both top line and bottom line; even open source has led to sticker shock on occasion; the industry did not have a vision of how networking was going to be reimagined; CAPEX and OPEX budget silos create commercial and transformation challenges.
Transformation fails across industries
Most transformation efforts fail, in most companies, regardless of industry. The long pole in the tent is people. People who have been doing the same thing, by habit, for decades; people who have been told for decades that doing XYZ makes them valuable and adds value to the company and its customers; people that are reluctant to change; people that focus on short-term incentives; people that focus on functional incentives within a corporation. Specifically, common reasons for failure include leadership alignment, lack of resources, lack of a communicated compelling vision, lack of engagement by leaders at the needed layer of the organization, complexity of reinventing the value chain and its processes, and culture. For more, read transformation links.
The net of the above is that for any large and/or old company, organic transformation is extremely hard. Inorganic and incubation strategies should be given a high priority to overcome the predictable and common challenges with organic change. I would argue that some of the specific NFV problems cited, may not exist, if NFV had been incubated in truly clean-slate organizations / acquisitions. Truly clean-slate means truly isolated geographically and virtually from the headquarters / mainstream company culture. Lastly, when inorganic strategies are executed, consideration needs to be given to putting the acquired entity in charge of the existing organizational structure. That is a complex subject and probably is a “depends” situation. However, it should be a consideration.
NFV Specific Challenges
This section most directly responds to points Tom articulated as the Operator View.
High licensing fees
Suppliers have to manage top line revenue as well as bottom line/profit. A company could make a choice just to manage for profit, but that risks how the street will react to revenue decline, and necessitates a significant transformation communication effort. Which leads to chicken and egg problem – why take this journey until it is clear the market is there (innovators dilemma). There is some truth to the idea that suppliers have significant OPEX costs related to BOTH R&D and S&M. This is clear from any public P&L.
There is much more that could be said about licensing fees, that would be a post in and of itself. But I would net it as aligning value proposition to licensing approach.
It is not clear to me that Open Source is necessarily significantly less expensive, especially when total life cycle costs are considered. You probably need support for any Open Source offering so there will be that cost at least. In theory that should be lower than the product+service cost of a non Open Source offering, but I’m not sure that is a given. The economics of Open Source is impacted not only by those entities supplying/supporting Open Source offerings, but also by the operational culture of the buyer. I would say the real benefit of Open Source would be a community taking a clean slate approach to reimagining networking. So Open Source could still be powerful, even if the economics is not always predictable. Note: operators have historically experienced sticker shock to the cost of some open source offerings (this was some years ago).
I would agree that API-based approaches is the way to go. But APIs for what? Another chicken / egg problem. Was there an industry vision that provided a North Star for how networking was going to be reimagined? And if not, is that why the box-paradigm was simply reproduced? (Yes, the box-paradigm was produced partly due to all the people challenges cited above).
NFV benefits vs investment required
Wasn’t DSL unprofitable at first until it was operationalized? Maybe the issue here is the CFO just doesn’t see the opportunity. I’m not saying there is no opportunity, but maybe the CFO has not seen it. I believe Tom, like others, has said publicly that SD-WAN may be the biggest real world driver(?) for NFV. Notwithstanding there are some in the SD-WAN world that are not aligned to NFV as a need in SD-WAN, the bigger point is the SD-WAN value prop is understood, has momentum, and therefore, there is line of sight to the opportunity. IF SD-WAN pulls through NFV, it is not (just) because NFV is intrinsically good, but because SD-WAN has a compelling value proposition and NFV is a way to realize it.
CAPEX vs OPEX
CAPEX and OPEX budgets are silo’d at many large operators. This makes it difficult to address costs holistically, and creates difficulty in constructing commercials. Some of the NFV evangelists at major CSPs had a vision of shift to OPEX, but that was killed by the amount of internal change / culture transformation that required. This speaks to transformation dynamics independent of industry, even if the silo itself is more typical of large Capex-intensive industries. Another important OPEX question is whether operational excellence in a clean-slate NFV world would decrease overall OPEX costs.
NFV is too complex
I cannot speak in detail to standards work, however, the bigger issue was probably lack of industry vision and lack of clean-slate implementations aligned to that vision. There are many reasons for complexity including design by committee, but another is a concern with maintaining existing capability while trying to push in new directions.